I have been struggling with getting consistently successful runs of my test lists using the scheduling and execution servers. Here are the details:
The test lists in question include several scripts which all follow the same procedure:
1. Navigate to and open application
2. Log in to application
3. Either look up existing "project" or create new one depending on test script
4. Open particular page of "project"
5. Exercise all options on the page
All scripting errors have been resolved in quick execution mode. There are delays placed strategically throughout the scripts and stepWaitOnElements is used. I have spent literally weeks on getting the scripts to this point.
When running the test lists remotely, test scripts either:
1. Die on the navigation step
2. Timeout on step three. There is a execution delay prior to this step with the time set to 30000 or 60000 msec depending on which application is being tested.
3. Run successfully.
Failing on navigation does not occur consistently as you will see in the attached list. It is not all at the beginning of the test list nor all at the end, but interspersed among test scripts. This proves true on repeated runs too, just which test scripts fail changes. The attached list has annotations of "L" (look up project) or "N" (new project) to indicate what is happening with step 3 above. I was just trying to find a pattern.
Some other notes related to resource issues: 1) The server is dedicated to Test Studio with myself and one other person using it. 2) The memory has been upped to 8GB. 3) Consistent clearing of browser cache is done. 4) Location of test script failure is not consistent. 5) If I reboot the server, generally the first test script executed will run without error.
Any ideas of where I should go from here to make these test lists less brittle?
Your help is always appreciated! Thanks!